Hi, I'm Anshuman
CS Undergrad at IIIT Kottayam & Lead, App Development Subclub at BetaLabs.
Previously AI Engineering Intern, building multi-agent analytics frameworks in LangGraph, production LLM evaluation pipelines, and observability infrastructure.
I enjoy building agentic AI infrastructure, distributed systems, and storage engines — exploring how systems work under the hood, from low-level LSM tree internals to real-time event streaming pipelines.
-ED2224?style=for-the-badge&logo=adobeacrobatreader&logoColor=white){kind=icon}
- AI & Agent Infrastructure: LangGraph, LangChain, DeepAgents, supervisor agent orchestration, evaluation harnesses, deterministic tool execution
- Distributed Systems & Concurrency: Multithreading, atomic primitives, race condition diagnosis, distributed locking, idempotency guarantees
- Storage Engine Internals & Data Engineering: LSM Trees (WAL, MemTable, SSTables, Bloom Filters), write amplification trade-offs, Change Data Capture (CDC), zero-loss replication
- High-Performance Backend & Networking: FastAPI, REST APIs, WebSockets, connection pooling, backpressure handling, rate limiting
- Reliability & System Design: Circuit breakers, exponential backoff with jitter, dead-letter queues (DLQ), telemetry, p95/p99 latency profiling
|
Enterprise CDC Streaming Pipeline
|
Embedded LSM-Tree Storage Engine
|
Agentic Market & Trading Assistant
|
|
Core Developer Infrastructure & Release Tooling
|
Quantum Algorithms & Compiler Benchmarking (MQT Bench)
|
|
Agent Capability Authorization & Testing Infrastructure
|
Security Hardening & Log Sanitization Infrastructure
|
- Concurrency & Thread Safety: Mutexes, condition variables, lock contention analysis, atomic operations, race-condition mitigation, and memory lifecycle management.
- Distributed Reliability & Fault Tolerance: Idempotent message consumers, distributed rate-limiting (Token Bucket / Leaky Bucket), retry storm protection with exponential backoff & jitter, and Dead-Letter Queue (DLQ) patterns.
- Storage Mechanics & Access Patterns: LSM-tree design (sequential append-only logging vs random read penalties), in-memory MemTables, immutable SSTables, probabilistic membership testing with Bloom Filters, and atomic Lua-scripted cache transactions.
- Network Transport & API Design: Connection pooling, HTTP/REST state contracts, persistent WebSocket streaming, backpressure control, and payload serialization optimization.
- AI & Agent Architecture: LangGraph, LangChain, DeepAgents, Multi-Agent Systems, LLM Evaluation, Observability, RAG
- Core Systems & Concurrency: Multithreading, Thread Pools, Mutexes & Locks, Race Condition Triage, Memory Management (RAII, Pointers)
- Distributed Systems & Streaming: Apache Kafka, Debezium (CDC), Redis, Elasticsearch, Idempotency, Circuit Breakers, Dead-Letter Queues (DLQ)
- Storage Engines & Databases: LSM Trees, Write-Ahead Logs (WAL), SSTables, Bloom Filters, PostgreSQL, MySQL, Redis (Lua Scripting)
- Backend & Networking: FastAPI, Node.js, Express, REST APIs, WebSockets, Rate Limiting, Connection Pooling, Supabase
- Infrastructure & Tools: Docker, Kubernetes, Linux (POSIX, Bash), GitHub, CMake, Google Benchmark, VS Code
